betaFeedback

WorksPhysics and Astronomy

Substrate-induced spin-torque-like signal in spin-torque ferromagnetic resonance measurement

Abstract

Oxide thin films and interfaces with strong spin-orbit coupling have recently shown exceptionally high charge-to-spin conversion, making them potential spin-source materials for spintronics. Epitaxial strain engineering using oxide substrates with different lattice constants and symmetries has emerged as a mean to further enhance charge-to-spin conversion. However, high relative permittivity and dielectric loss of commonly used oxide substrates, such as SrTiO3, can cause significant current shunting in substrates at high frequency, which may strongly affect spin-torque measurement and potentially result in an inaccurate estimation of charge-to-spin conversion efficiency. In this study, we systematically evaluate the influence of various oxide substrates for the widely-used spin-torque ferromagnetic resonance (ST-FMR) measurement. Surprisingly, we observed substantial spin-torque signals in samples comprising only ferromagnetic metal on oxide substrates with high relative permittivity (e.g., SrTiO3 and KTaO3), where negligible signal should be initially expected. Notably, this unexpected signal shows a strong correlation with the capacitive reactance of oxide substrates and the leakage radio frequency (RF) current within the substrate. By revising the conventional ST-FMR analysis model, we attribute this phenomenon to a 90-degree phase difference between the RF current flowing in the metal layer and in the substrate. We suggest that extra attention should be paid during the ST-FMR measurements, as this artifact could dominate over the real spin-orbit torque signal from high-resistivity spin-source materials grown on substrate with high relative permittivity.

Introduction

Magnetic random-access memories (MRAMs) and spintronic logical devices based on spin-orbit torques (SOTs) have garnered significant research attention due to their low-power consumption, high endurance, high-speed, and non-volatile properties [1, 2, 3, 4, 5, 6, 7, 8, 9, 10]. The development of materials capable of generating SOT with high-efficiency and the establishment of accurate metrologies for quantifying current-induced SOTs in various material systems are crucial.

To evaluate SOT efficiency of spin source materials, several experimental techniques have been devised to detect static or dynamic changes in magnetization of ferromagnet (FM) induced by SOTs [11]. These methods include spin-torque ferromagnetic resonance (ST-FMR) [12, 13, 14, 15, 16, 17], second harmonic Hall voltage (SHHV) [18, 19, 20], measurements of current-induced hysteresis loop shift [21, 22, 23], measurements of current-induced magnetization switching [24], and optical measurements of current-induced effective field [25, 26, 27], etc. Among these techniques, ST-FMR stands out as a widely employed method for measuring SOTs due to its self-calibration, simplicity, and versatility across a wide range of material systems [12, 13, 14, 28, 29, 30, 31]. ST-FMR allows the examination of magnetization dynamics induced by SOTs at ferromagnetic resonance frequencies of the gigahertz range, providing insights into both damping-like and field-like SOTs. During the ST-FMR measurements, spin pumping from the magnetic layer to the spin source layer, coupled with the inverse spin-Hall effect, can contribute to the spin torque signal and is considered as the major source of artifact, which has been carefully evaluated [31, 32]. On the other hand, despite considerable attention paid to the fact that substrates could become highly conductive at high frequencies and generate spin-torque-like signals, the impact of substrates to the ST-FMR measurements has not been thoroughly investigated.

Remarkably, the influence of substrates could be even more pronounced when measuring SOTs in oxide spin source materials, which have recently exhibited an exceptionally strong spin-Hall effect [28, 30, 33, 34, 35, 36]. Firstly, oxide spin source materials typically have higher resistivity compared to heavy metals at room-temperature (i.e. Pt ∼20\sim20 μΩ\mu\Omegacm [12], SrIrO3_3 ∼570\sim570 μΩ\mu\Omegacm [37], SrRuO3_3 ∼290\sim290 μΩ\mu\Omegacm [36]). Secondly, oxide substrates, such as SrTiO3_3 and KTaO3_3 which are commonly employed for epitaxial growth of oxide spin sources [15, 29, 30, 35, 36] and two-dimensional electron gas (2DEG) systems [33, 38, 39, 40], exhibit high conductivity at high frequencies due to their high dielectric constant and dielectric loss [41, 42]. This combination leads to a much more pronounced RF current shunting effect from the substrate in oxide spin sources as compared to heavy metal materials in ST-FMR measurements.

In this work, we investigated the artifact spin-torque signal in ST-FMR measurements on different oxide substrates with a wide range of relative permittivity. We find that the artifact spin torque signal is large in samples of only ferromagnetic metal permalloy (Py) on oxide substrates with high relative permittivity (without a spin source material). This artifact spin torque signal is inversely proportional to the capacitive reactance of substrates, which can be attributed to the Oersted field generated by an off-phased leakage RF current in oxide substrates according to our revised ST-FMR analysis model. Notably, we demonstrate that this artifact spin-torque signal can significantly impact the spin-Hall ratio analysis of Pt, a benchmarked spin-Hall source material with high conductivity, when using SrTiO3_3 substrate. When using the more resistive oxide spin-Hall source materials such as SrIrO3_3, the spin torque signal from SrIrO3_3 could be overwhelmed by the artifact from SrTiO3_3 alone. Our findings offer a deeper insight into the origin of spin torque signal measured in oxide spin-Hall sources and pave the way for a more precise measurement of SOT materials.

Spin-torque ferromagnetic resonance (ST-FMR)

In ST-FMR measurements, a microwave RF current is applied on the microstrip of spin-Hall source material/FM bilayer. Due to the spin-Hall effect (SHE) [43] or Rashba-Edelstein effect (REE) [44], an RF spin current is generated in the spin-Hall source layer and inject into the adjacent FM layer to induce a magnetization precession, which leads to an oscillation of the microstrip resistance because of the anisotropic or spin-Hall magnetoresistance in the FM layer. The ST-FMR voltage signal can be detected by mixing the RF current with the oscillating resistance [12, 13, 45].

By solving the Landau-Lifshitz-Gilbert (LLG) equation, one can obtain the mixing voltage as a function of the external magnetic field [12, 13, 31], as expressed in Eq. (1):

Vmix(Hext)=VSΔH2ΔH2+(Hext−H0)2+VAΔH(Hext−H0)ΔH2+(Hext−H0)2(1)V_{\mathrm{mix}}(H_{\mathrm{ext}})=V_{\mathrm{S}}\frac{\Delta H^{2}}{\Delta H^{2}+(H_{\mathrm{ext}}-H_{0})^{2}}+V_{\mathrm{A}}\frac{\Delta H(H_{\mathrm{ext}}-H_{0})}{\Delta H^{2}+(H_{\mathrm{ext}}-H_{0})^{2}} \tag*{(1)}

where HextH_{\mathrm{ext}} is the external magnetic field, ΔH\Delta H and H0H_{0} are the linewidth and resonance field, respectively. By employing the fitting procedure of Eq. (1) consisting of symmetric and antisymmetric Lorentzian functions to the experimental data, two voltage components can be extracted: a symmetric voltage component VSV_{\mathrm{S}} proportional to the strength of the damping-like torque τDL\tau_{\mathrm{DL}}, and an antisymmetric voltage component VAV_{\mathrm{A}} proportional to the strength of the sum of field-like torque τFL\tau_{\mathrm{FL}} and RF Oersted field HRFH_{\mathrm{RF}}, according to the Eq. (2) and Eq. (3), respectively:

VS=−14dRdφIRFcos⁡φ2πΔH(dfdHext∣Hext=H0)τDL(2)V_{\mathrm{S}}=-\frac{1}{4}\frac{dR}{d\varphi}\frac{I_{\mathrm{RF}}\cos\varphi}{2\pi\Delta H\left(\left.\frac{df}{dH_{\mathrm{ext}}}\right|_{H_{\mathrm{ext}}=H_{0}}\right)}\tau_{\mathrm{DL}} \tag*{(2)}
VA=−14dRdφIRFcos⁡φ1+(4πMeff/H0)2πΔH(dfdHext∣Hext=H0)(τFL+γμ0HRF)(3)V_{\mathrm{A}}=-\frac{1}{4}\frac{dR}{d\varphi}\frac{I_{\mathrm{RF}}\cos\varphi\sqrt{1+\left(4\pi M_{\mathrm{eff}}/H_{0}\right)}}{2\pi\Delta H\left(\left.\frac{df}{dH_{\mathrm{ext}}}\right|_{H_{\mathrm{ext}}=H_{0}}\right)}\left(\tau_{\mathrm{FL}}+\gamma\mu_{0}H_{\mathrm{RF}}\right) \tag*{(3)}

where ff is the excitation frequency of ST-FMR measurement, γ\gamma is the gyromagnetic ratio, μ0\mu_{0} is the permeability in vacuum, RR is the resistance of microstrip, IRFI_{\mathrm{RF}} is the RF current through the microstrip, 4πMeff4\pi M_{\mathrm{eff}} is the effective demagnetization field, φ\varphi is the angle between IRFI_{\mathrm{RF}} and HextH_{\mathrm{ext}}. Finally, the spin-Hall ratio θSH\theta_{\mathrm{SH}} can be calculated via [15, 45]:

θSH=τDL2eℏMstlρIRFR(4)\theta_{\mathrm{SH}}=\tau_{\mathrm{DL}}\frac{2e}{\hbar}\frac{M_{\mathrm{s}}tl\rho}{I_{\mathrm{RF}}R} \tag*{(4)}

where ℏ\hbar is the reduced Planck’s constant, ee is the electron charge, ll is the length of the microstrip, tt is the thickness of FM layer, ρ\rho is the resistivity of spin-Hall source layer. In Eq. (4), the only unknown parameter is the RF current IRFI_{\mathrm{RF}} which can be measured using a vector network analyzer [45, 46]. Alternatively, in a simpler scenario, the spin-Hall ratio θSH\theta_{\mathrm{SH}} can be estimated by considering the ratio of VSV_{\mathrm{S}} to VAV_{\mathrm{A}}, denoted as

θSH=(VSVA)1+4πMeffH0eμ0Mstdℏ\theta_{\mathrm{SH}}=\left(\frac{V_S}{V_A}\right)\sqrt{1+\frac{4\pi M_{\mathrm{eff}}}{H_0}}\frac{e\mu_0 M_s t d}{\hbar}

assuming that VAV_A is primarily contributed from the Oersted field, where dd is the thickness of spin-Hall source layer [12, 29, 30]. It should be noted that in systems with significant field-like torques [17, 47, 48], the spin-Hall ratio obtained solely from the ratio of VSV_S to VAV_A may lead to misestimations. Nevertheless, accurate measurements of the symmetric component (VSV_S) are of great importance for determination of the spin-Hall ratio in ST-FMR measurements.

Device preparation and measurements

In this study, we used oxide substrates with different relative permittivity εr\varepsilon_r at room temperature during the ST-FMR measurements. The substrates used were 5 mm×5 mm, 500 μm-thickness (001)-oriented SrTiO3_3 (STO, εr∼300\varepsilon_r \sim300[41]), (001)-oriented KTaO3_3 (KTO, εr∼290\varepsilon_r \sim290[42]), (001)-oriented (LaAlO)0.3_{0.3}-(Sr2_2AlTaO6_6)0.7_{0.7} (LSAT, εr∼22.6\varepsilon_r \sim22.6[50]), (110)-oriented DyScO3_3 (DSO, εr∼26\varepsilon_r \sim26[51]), and (0001)-oriented Al2_2O3_3 (εr∼9.2\varepsilon_r \sim9.2[52]), and all substrates were not be treated with any solutions. To make the samples without any spin-Hall source materials, we deposited only the ferromagnetic thin films of Py (5 nm) capped with Ti (1.5 nm) onto these oxide substrates by using high-vacuum magnetron sputtering at an Ar pressure of 3 mTorr. To make the control samples, bilayers of Pt (5nm)/Py (5nm) and Py (5nm)/Pt (5nm) with reversed layered sequence were deposited on STO, and Pt (5nm)/Py (5nm) was deposited on Si (ρ∼0.01\rho\sim0.01 Ω ∙ cm)/SiO2_2 (500 nm). For samples with oxide spin-Hall-source materials, 17 nm SrIrO3_3 (SIO) was grown on a (001)-oriented STO substrate by using a pulsed laser deposition (PLD) with a 248 nm KrF excimer laser at an oxygen pressure of 0.1 Torr. During the PLD growth, the growth temperature was 700 °C, and the laser energy density was maintained at 1.2 J/cm2^2. Subsequently, the sample was cooled to room temperature in an oxygen atmosphere of 30 Torr. Then a 5 nm Py was deposited onto the samples by sputtering. All samples were fabricated using the lithography and ion-beam etching processes, into the 16 μm×80 μm microstripes with ground-signal-ground (GSG) electrodes for ST-FMR measurements and 16 μm×80 μm Hall bars for SHHV measurements. Especially, for the STO/Py, STO/Pt/Py and STO/Py/Pt samples, we confirmed that the STO substrates were still insulative after the device fabrication processes by measuring the dc resistance (> 200 MΩ) between two metal pads separated by ∼ 200 μm.

Figure 1 shows the schematic of ST-FMR measurement setup, in which a microwave current with a power of 15 dBm is applied on the microstrip along the +x+x direction from a microwave generator and is modulated by a low frequency sinewave at 1713 Hz. The mixing voltage signal was measured by a lock-in amplifier as a function of the external magnetic field HextH_{ext}. To characterize the impedance of samples, we used a vector network analyzer (VNA) to measure the scattering parameter S11S_{11} (see S1 in Supplementary Material [49] for a detailed description of the measurements).

Schematic of ST-FMR measurement set-up and RF current shunting model

Figure 1. Schematic of ST-FMR measurement set-up and RF current shunting model. HextH_{ext} is the external magnetic field, φ\varphi is the angle between RF current IRFI_{RF} and external magnetic field HextH_{ext}, mm is the magnetization of Py layer. IRFI_{RF}, IRF−substrateI_{RF-substrate}, IRF−PyI_{RF-Py} are the total RF current through oxide substrate/Py device, leakage RF current in oxide substrate, and RF current in Py, respectively. RPyR_{Py} is the equivalent resistor of Py layer, and CsubstrateC_{substrate} is the capacitor formed in oxide substrate through the ground-signal-ground (GSG) electrode.

Results

Figure 2(a)-(Figure 2e) show the typical ST-FMR spectra of 5 nm Py on substrates of STO, KTO, Al2_2O3_3, DSO, and LSAT measured at f=5f = 5 GHz, φ=225∘\varphi= 225^\circ, which can be well fit by the sum of symmetric and antisymmetric Lorentzian functions expressed in Eq. (1). We found that the symmetric voltage signal that was attributed to damping-like torque appears in all five samples. Noticeably, the symmetric voltage signal from STO/Py (5nm), KTO/Py (5nm) has the same order of magnitude as Si/SiO2_2/Pt (5nm)/Py (5nm) control sample which has a substantial damping-like torque with a spin-Hall ratio of 0.071 ±\pm 0.002 (see S2 in Supplementary Material [49] for details). However, in all samples including STO/Py and KTO/Py, the damping-like torque should be negligible since there is no spin source layer [12], as confirmed by an independent low-frequency SHHV measurements (see S3 in Supplementary Material [49] for details of SHHV measurements). The discrepancy between ST-FMR and SHHV measurements for the oxide substrate/Py devices (especially the heterostructures of STO/Py and KTO/Py) implies that the symmetric signal from ST-FMR measurements is not originated from SOT but might be an artifact signal. Furthermore, we discovered that this artifact signal possessed the same symmetry as that of SOT, with the same angular dependence as Si/SiO2_2/Pt/Py control sample (see Figure S6 in Supplementary Material [49]). Additionally, small antisymmetric components are also observed from all oxide substrate/Py samples, which could be originated from a field-like torque due to the Rashba effect at the Py/Ti interface [53].

The typical ST-FMR spectra of STO/Py, KTO/Py, Al$_2$O$_3$/Py, DSO/Py, LSAT/Py, and Si/SiO$_2$/Pt/Py

Figure 2. The typical ST-FMR spectra of (a) STO, (b) KTO, (c) Al2_2O3_3, (d) DSO, (e) LSAT oxide substrate/Py (5nm)

Angular dependence of symmetric and antisymmetric components for six oxide substrate/Py devices

Figure S6. Angular dependence of symmetric components and antisymmetric components of (a)-(e) different oxide substrate/Py devices, and (f) Si/SiO2_2/Pt (5nm)/Py (5nm) device, which can be well fit by sin⁡(2φ)cos⁡(φ)\sin(2\varphi)\cos(\varphi).

devices, and (f) Si/SiO2_2/Pt (5 nm)/Py (5 nm) device at f=5f = 5 GHz, φ=225∘\varphi= 225^\circ, P=15P = 15 dBm. The black hollow squares represent the experimental data. The black lines, red lines, and blue lines are the sum of symmetric and antisymmetric Lorentz functions fitting VmixV_{\mathrm{mix}}, symmetric Lorentzian function fitting VSV_{\mathrm{S}}, and antisymmetric Lorentzian function fitting VAV_{\mathrm{A}}, respectively.

To understand the origin of the artifact symmetric voltage signal observed in ST-FMR measurements, we established a revised ST-FMR analysis model by considering the leakage RF current in oxide substrates. As shown in Figure 1, the device can be considered as a resistance RPyR_{\mathrm{Py}} of Py and a capacitor CsubstrateC_{\mathrm{substrate}} formed between two electrodes that are connected in parallel. At the frequency of ST-FMR measurement (5-10 GHz), the capacitive reactance of CsubstrateC_{\mathrm{substrate}} decreases as the frequency increases, and thereby the substrates might become conductive at radio frequencies. As a result, part of the RF current will flow through the oxide substrate due to the current shunting. Therefore, as shown in Figure 1, the total RF current IRFI_{\mathrm{RF}} through the oxide substrate/Py device consists of two components: the RF current IRF−PyI_{\mathrm{RF-Py}} through RPyR_{\mathrm{Py}}, and the RF current IRF−substrateI_{\mathrm{RF-substrate}} through CsubstrateC_{\mathrm{substrate}}. Similar to the conventional spin-Hall source material/FM bilayer, IRF−substrateI_{\mathrm{RF-substrate}} in the oxide substrate will generate an RF Oersted field HRF′H'_{\mathrm{RF}}, as shown in Figure 3(a). However, the field-like torque of Oersted field would generally give rise to an antisymmetric signal [12], which cannot explain the artifact symmetric signal we observed. It should be noted that there is a 90-degree phase difference between IRF−PyI_{\mathrm{RF-Py}} and IRF−substrateI_{\mathrm{RF-substrate}}, as shown in Figure 3(b). Thereby, due to the 90-degree phase difference between the symmetric and antisymmetric signals, the off-phased HRF′H'_{\mathrm{RF}} generated from IRF−substrateI_{\mathrm{RF-substrate}} would cause the appearance of the artifact symmetric signal.

Sectional view of the oxide substrate/Py device, RF-current relation, impedance measurements, and correlation between artifact symmetric voltage and capacitive reactance

Figure 3. (a) Sectional view of the oxide substrate/Py device. HRF′H^{\prime}_{RF} is the RF Oersted field generated by the leakage RF current IRF-substrateI_{RF\text{-}substrate} in oxide substrate. (b) The relation between total RF current IRFI_{RF}, RF current in oxide substrate IRF-substrateI_{RF\text{-}substrate}, and RF current in Py IRF-PyI_{RF\text{-}Py}. The length of the arrow represents the modulus of RF current and αdevice\alpha_{\mathrm{device}} is the phase of impedance of the oxide substrate/Py device. (c) The modulus of impedance ∣Z∣\lvert Z\rvert of the different oxide substrates, oxide substrate/Py devices, and the phase of impedance αdevice\alpha_{\mathrm{device}} of the different oxide substrate/Py devices measured using a vector network analyzer at f=5f=5 GHz, respectively. (d) The correlation between the absolute value of the artifact symmetric voltage signal VS′V^{\prime}_{S} and the capacitive reactance of oxide substrate

To account for this signal, we revised the ST-FMR analysis model starting with the LLG equation to get the mixing voltage of oxide substrate/Py (see S4 in Supplementary Material [49] for details). The revised mixing voltage is then expressed as Eq. (5):

Vmix′=−14dRdφ∣IRF∣cos⁡φΔH2πdfdHext∣Hext=H01+4πMeffH0γμ0∣IRF−substrate∣2wFS(Hext)=VS′FS(Hext)(5)\begin{aligned} V'_{\mathrm{mix}}=-\frac{1}{4}\frac{dR}{d\varphi}\frac{\lvert I_{\mathrm{RF}}\rvert\cos\varphi}{\Delta H\left.2\pi\frac{df}{dH_{\mathrm{ext}}}\right|_{H_{\mathrm{ext}}=H_0}}\sqrt{1+\frac{4\pi M_{\mathrm{eff}}}{H_0}}\frac{\gamma\mu_0\lvert I_{\mathrm{RF-substrate}}\rvert}{2w}F_{\mathrm{S}}\left(H_{\mathrm{ext}}\right) \\ &=V'_{\mathrm{S}}F_{\mathrm{S}}\left(H_{\mathrm{ext}}\right) \tag*{(5)} \end{aligned}

where ww is the width of micro stripe, FS(Hext)F_{\mathrm{S}}\left(H_{\mathrm{ext}}\right) is the symmetric Lorentzian function, VS′V'_{\mathrm{S}} is the amplitude for that symmetric signal. From Eq. (5), we could conclude that the field-like torque of HRF′∝IRF,substrateH'_{RF} \propto I_{RF,\mathrm{substrate}} would actually give rise to an artifact symmetric signal Vs′V'_{s} in the revised model owing to the 90-degree phase difference between IRF−substrateI_{RF-\mathrm{substrate}} and IRF−PyI_{RF-\mathrm{Py}}. As we mentioned earlier, the symmetric signal can only be generated when there is a finite damping-like torque from the spin-Hall source layer in the conventional ST-FMR analysis model (Eq. (1) and Eq. (2)). To link the artifact symmetric signal to the capacitive reactance of substrate, we rewrite Vs′V'_{s} as:

Vs′=[−14dRdφγμ0∣P∣cos⁡φΔH 2πdf/dHext∣Hext=H0(1+(4πMeff/H0)2−12w)]1∣χsubstrate∣(6)V'_{s}=\left[-\frac{1}{4}\frac{dR}{d\varphi}\frac{\gamma\mu_{0}|P|\cos\varphi}{\Delta H\,2\pi df/dH_{\mathrm{ext}}\mid_{H_{\mathrm{ext}}=H_{0}}}\left(\sqrt{1+\left(4\pi M_{\mathrm{eff}}/H_{0}\right)^{2}}-\frac{1}{2w}\right)\right]\frac{1}{|\chi_{\mathrm{substrate}}|} \tag*{(6)}

where ∣P∣|P| is the modulus of power of the microwave used in ST-FMR measurements which can be expressed as ∣P∣=∣IRF∣2∣Zdevice∣|P|=|I_{RF}|^{2}|Z_{\mathrm{device}}|, and χsubstrate\chi_{\mathrm{substrate}} is the reactance of CsubstrateC_{\mathrm{substrate}} which can be expressed as χsubstrate=1/j2πfCsubstrate\chi_{\mathrm{substrate}}=1/j2\pi fC_{\mathrm{substrate}}. The part on the left side of 1/∣χsubstrate∣1/|\chi_{\mathrm{substrate}}| in Eq. (6) can be approximatively regarded as a constant and thereby Vs′V'_{s} would show a linear dependence on 1/∣χsubstrate∣1/|\chi_{\mathrm{substrate}}|.

To experimentally explore the above correlation, we have first measured the impedance of oxide substrate/Py devices and bare oxide substrates by using a network analyzer through the GSG electrodes at frequencies of 5-10 GHz. Due to the parallel connection, the impedance of oxide substrate/Py device ZdeviceZ_{\mathrm{device}} can be expressed as Eq. (7):

Zdevice=RPy1+(2πfCsubstrateRPy)2(1−j2πfCsubstrateRPy)(7)Z_{\mathrm{device}}=\frac{R_{\mathrm{Py}}}{1+\left(2\pi f C_{\mathrm{substrate}}R_{\mathrm{Py}}\right)^{2}}\left(1-j2\pi fC_{\mathrm{substrate}}R_{\mathrm{Py}}\right) \tag*{(7)}

Figure 3(c) shows the modulus of impedance for oxide substrate/Py devices and bare oxide substrates, along with the phase of impedance of the oxide substrate/Py devices. We found that ∣Zsubstrate∣|Z_{\mathrm{substrate}}| (∣χsubstrate∣|\chi_{\mathrm{substrate}}|) decreases as Csubstrate∝εrC_{\mathrm{substrate}}\propto\varepsilon_{r} and ∣Zdevice∣|Z_{\mathrm{device}}| is smaller than ∣Zsubstrate∣|Z_{\mathrm{substrate}}| for different oxide substrates, and αdevice\alpha_{\mathrm{device}} indicated by αdevice=arctan⁡(−2πf⋅RPyCsubstrate)\alpha_{\mathrm{device}}=\arctan(-2\pi f\cdot R_{\mathrm{Py}}C_{\mathrm{substrate}}) being close to -90 degrees as ∣Zsubstrate∣|Z_{\mathrm{substrate}}| decreases, which indicates the parrellel connection between RPyR_{\mathrm{Py}} and CsubstrateC_{\mathrm{substrate}}. Especially, ∣Zdevice∣|Z_{\mathrm{device}}| becomes close to ∣Zsubstrate∣|Z_{\mathrm{substrate}}| for STO and KTO due to their high relative permittivity, suggesting that IRF−substrateI_{RF-\mathrm{substrate}} would dominant over IRF−PyI_{RF-\mathrm{Py}} in STO/Py and KTO/Py devices. Figure 3(d) summarizes the measured artifact symmetric signals as a function of 1/∣χsubstrate∣1/|\chi_{\mathrm{substrate}}| for different oxide substrate/Py devices at a frequency of 5 GHz (see Figure S13 in Supplementary Material [49] for the correlations between VS′V^{\prime}_{S} and 1/∣χsubstrate∣1/\lvert\chi_{\mathrm{substrate}}\rvert at 6–10 GHz). The deviation from the linear fitting of VS′V^{\prime}_{S} v.s. 1/∣χsubstrate∣1/\lvert\chi_{\mathrm{substrate}}\rvert for samples with small permittivity substrates at higher frequencies shown in Figure S13 may be attributed to some self-induced spin-torque signals from single Py and other spurious effects such as the spin pumping from Py [54, 55]. As shown in Figure 2, given the significant dependence on substrates, other self-induced effects of the single Py are negligible compared with the artifact symmetric signal arising from the leakage RF current in substrate. The observed linear relation between VS′V^{\prime}_{S} and 1/∣χsubstrate∣1/\lvert\chi_{\mathrm{substrate}}\rvert is in good agreement with our model expressed as Eq. (6), providing strong evidence that the artifact symmetric voltage signal in oxide substrate/Py indeed arises from the leakage RF current in the oxide substrate.

Five-panel scatter plots showing the linear correlation between artifact symmetric signal and inverse substrate susceptibility

Figure S13. The linear correlation between the absolute value of artifact symmetric signal ∣VS′∣\lvert V'_{S}\rvert and 1/∣χsubstrate∣1/\lvert\chi_{\mathrm{substrate}}\rvert at frequencies of 6-10 GHz, φ=225∘\varphi= 225^\circ. The dashed line is the linear fit.

∣χsubstrate∣\left|\chi_{\mathrm{substrate}}\right| at f=5GHzf = 5\mathrm{GHz}, φ=225∘\varphi= 225^\circ. The dashed line is a linear fit.

To directly evaluated the influence of the artifact symmetric signal from oxide substrates with high relative permittivity on determining the spin-Hall ratio with the ST-FMR measurements, we prepared Pt (5nm)/Py (5nm) and Py (5nm)/Pt (5nm) with a reversed layered sequence on the STO substrate. In these samples, the total RF current IRFI_{\mathrm{RF}} consists of IRF−PyI_{\mathrm{RF-Py}}, IRF−PtI_{\mathrm{RF-Pt}} and IRF−substrateI_{\mathrm{RF-substrate}}, in which IRF−PtI_{\mathrm{RF-Pt}} gives rise to a symmetric signal VSV_{\mathrm{S}} from damping-like torque on Py, while IRF−substrateI_{\mathrm{RF-substrate}} gives rise to the artifact symmetric signal VS′V_{\mathrm{S}}^{\prime}. As shown in Figure 2, the artifact signal from IRF−substrateI_{\mathrm{RF-substrate}} has the same symmetry as that from IRF−PtI_{\mathrm{RF-Pt}}, which means the VS′V_{\mathrm{S}}^{\prime} and VSV_{\mathrm{S}} have the same sign when Pt and oxide substrate is on the same side of Py. However, when Pt and oxide substrate are on the opposite side of Py, VS′V_{\mathrm{S}}^{\prime} and VSV_{\mathrm{S}} would have an opposite sign and cancel out to each other. Specifically, in the STO/Pt/Py sample, VS′V_{\mathrm{S}}^{\prime} from STO would add up to VSV_{\mathrm{S}} from Pt as shown in Figure 4(a), resulting in an overestimation of θSH\theta_{\mathrm{SH}} for Pt; while conversely, in the STO/Py/Pt sample, VS′V_{\mathrm{S}}^{\prime} would counteract VSV_{\mathrm{S}}, as shown in Figure 4(b), resulting in an underestimation of θSH\theta_{\mathrm{SH}}.

Schematic of the influence of the artifact symmetric voltage and comparison of ST-FMR and SHHV results

Figure 4. Schematic of the influence of the artifact symmetric voltage on the spin-Hall ratio of Pt in (a) STO/Pt/Py and (b) STO/Py/Pt samples. IRFI_{\mathrm{RF}} is the total RF current. IRF−PyI_{\mathrm{RF-Py}}, IRF−PtI_{\mathrm{RF-Pt}} and IRF−substrateI_{\mathrm{RF-substrate}} are the RF current in Py, Pt and STO substrate, respectively. VSV_{\mathrm{S}} is the symmetric component related to the damping-like torque from IRF−PtI_{\mathrm{RF-Pt}}, VS′V_{\mathrm{S}}^{\prime} is the artifact symmetric signal generated from the leakage RF current IRF−substrateI_{\mathrm{RF-substrate}} in STO. (c) The spin-Hall ratio θSH\theta_{\mathrm{SH}} obtained from ST-FMR measurements, along with the damping-like effective fields HDLH_{\mathrm{DL}} obtained from SHHV of STO/Pt/Py and STO/Py/Pt.

Figure 4(c) shows the spin-Hall ratio θSH\theta_{\mathrm{SH}} extracted from the ST-FMR measurements (see S5 in Supplementary Material [49] for details). θSH\theta_{\mathrm{SH}} is found to be 0.097 ±\pm 0.001 for STO/Pt/Py, which is about 2.5 times larger than 0.038 ±\pm 0.001 for STO/Py/Pt, consistent with our model. To validate our findings further, we measured the damping-like torque of STO/Pt/Py and STO/Py/Pt by using an independent low-frequency SHHV measurements (see S5 in Supplementary Material [49] for details). In this measurement, the leakage current in oxide substrate should be minimal since the driving frequency for the measurement is low (233 Hz). The damping-like effective fields HDLH_{\mathrm{DL}} for these two samples extracted from SHHV were found to be almost the same within experimental uncertainties, as shown in Figure 4(c). The large discrepancy of the spin-Hall ratios for these two samples using ST-FMR measurements (while almost no difference when using SHHV) demonstrates that substrates with high relative permittivity significantly influence the estimation of the spin-Hall ratio, even in spin-Hall source materials with high conductivity, in ST-FMR measurements.

Discussion

While the self-calibrated ST-FMR measurement offers unique advantages over other spin torque measurement techniques, it is crucial to recognize that certain substrates with high relative permittivity can introduce significant artifact spin-torque signals in ST-FMR measurement, particularly in spin-Hall source materials with high resistivity. For example, 4d/5d transition metal oxide thin films have recently exhibited exceptionally large spin-Hall ratios [15, 28, 29, 30, 34, 37], often possess higher resistivity than heavy metals. When measuring the oxide thin films grown on high-relative-permittivity substrates such as STO and KTO using ST-FMR measurement, their large resistivity can lead to the majority of RF current flowing in substrate, leading to a pronounced artifact symmetric signal.

Previous studies on 4d/5d transition metal oxide thin films and oxide 2DEG systems have frequently used the ST-FMR technique with the analysis of the symmetric signal. Large variations in amplitude and even in sign of spin-Hall ratio have been for the same spin-Hall material with different epitaxial substrates and layered structure [15, 28, 29, 30, 33, 37, 38, 56]. Our results indicate that this varation could mainly be attributed to the artifact symmetric signal caused by using substrates with high relative permittivity (Figure 3(d)) and the sequence of heterostructures (Figure 4(a) and 4(b)), rather than the other intrinsic mechanisms. For instance, our ST-FMR results for the STO/SIO (17nm)/Py (5nm) sample (see S6 in Supplementary Material [49] for details) show nearly the same magnitude of symmetric signal as compared to the STO/Py (5nm) sample (Figure 2(a)), indicating that the influence of artifact symmetric signal from STO is overwhelming.

We proposed that employing the dc current-tuned ST-FMR measurements [12, 13, 15, 29] may eliminate the influence of the artifact signal from substrate since there is no obvious correlation between the artifact symmetric signal and the current-induced change of linewidth (see Figure S7 in Supplementary Materials [49]). Furthermore, spin torque measurement techniques at quasi-static frequencies, such as SHHV [29, 57, 58] or current-induced loop-shift[21, 22, 23], should be adopted when characterizing the spin-Hall materials with high resistivity on substrate with high permittivity.

Kittle formula fitting and resonant linewidth of oxide substrate/Py devices

Figure S7. (a) The Kittle formula fitting of resonant frequency ff and resonant field H0H_0 of different oxide substrate/Py devices, and inset is the extracted effective demagnetization of Py in different oxide substrate/Py devices. (b) Resonant linewidth of different oxide substrate/Py devices at frequencies of 5-10 GHz, φ=225∘\varphi=225^\circ.

Conclusions

In conclusion, we have uncovered the presence of a significant artifact symmetric signal in ST-FMR measurements of oxide substrate/Py devices. By revising the conventional ST-FMR analysis model, we have demonstrated that the artifact symmetric signal arises from the leakage RF current in the oxide substrate, caused by the RF current shunting effect. We have further established a correlation between the magnitude of the artifact symmetric signal and the reactance of the capacitor formed in substrate through the GSG electrode. We have evaluated the influence of artifact symmetric signal on the spin-Hall ratio of Pt from oxide substrate with high relative permittivity (i.e., STO) by carrying out the ST-FMR and SHHV measurements, which indicates that the substrates with high relative permittivity significantly affect the evaluation of the spin-Hall ratio in the ST-FMR measurements. Our findings shed light on the critical influence of substrate-induced artifacts in ST-FMR measurements and highlight the importance of usage of suitable measurement techniques when studying spin-Hall source materials.

Acknowledgments

This work was supported by the National Natural Science Foundation (52161135103, 52250418), and Tsinghua University Initiative Scientific Research Program. W.S. acknowledges grant no. 2021/40/Q/ST5/00209 (Sheng) from the National Science Centre, Poland.

Artifact from Spin-Torque Ferromagnetic Resonance (ST-FMR) Measurements with Oxide Substrates

Dingsong Jiang; Hetian Chen; Guiping Ji; Yahong Chai; Chenye Zhang; Yuhan Liang; Jingchun Liu; Witold Skowroński; Pu Yu; Di Yi; Tianxiang Nan

#: The authors contribute equally to this work.

*: [email protected], [email protected]

The measurements of the impedance of different oxide substrates and oxide substrate/Py devices.

We used the vector network analyzer (VNA) to measure the S11S_{11} reflection parameters of bare oxide substrates and oxides substrate/Py devices through the ground-signal-ground (GSG) electrodes at frequencies of 1-15 GHz, as shown in Figure S1. Subsequently, the impedance of substrate ZsubstrateZ_{substrate} and substrate/Py device ZdeviceZ_{device} can be evaluated via

Schematic of using the VNA to measure reflection parameters of an oxide substrate and an oxide substrate/Py device through GSG electrodes

Figure S1. Schematic of using the VNA to measure the S11S_{11} reflection parameter of (a) oxide substrate and (b) oxide substrate/Py device through the ground-signal-ground (GSG) electrodes. When measuring the S11S_{11} reflection parameter of the bare oxide substrate, we only fabricated the GSG electrodes on the substrate.

Z=50Ω⋅1+S111−S11,Z = 50\Omega\cdot\frac{1+S_{11}}{1-S_{11}},

where ZZ and S11S_{11} are both complex numbers. We calculated the mean value of the 100 data points around each frequency to obtain the ZsubstrateZ_{substrate} and ZdeviceZ_{device} at 5, 6, 7, 8, 9, 10 GHz, as shown in Figure S2 and Figure S3.

Measured modulus and phase of impedance of different bare substrates

Figure S2. The measured modulus ∣Zsubstrate∣\lvert Z_{\mathrm{substrate}}\rvert and phase of impedance αsubstrate\alpha_{\mathrm{substrate}} of the different bare substrates at frequencies of 5-10 GHz. ∣Zsubstrate∣\lvert Z_{\mathrm{substrate}}\rvert decreases along with the frequency and αsubstrate\alpha_{\mathrm{substrate}} is around -90 degrees, which indicates that the capacitor is formed between the GSG electrodes.

Modulus and phase of impedance of different oxide substrates and oxide substrate/Py devices

Figure S3. The modulus of impedance of different oxide substrates ∣Zsubstrate∣\lvert Z_{\mathrm{substrate}}\rvert, oxide substrate/Py devices ∣Zdevice∣\lvert Z_{\mathrm{device}}\rvert and the phase of impedance of the different oxide substrate/Py devices αdevice\alpha_{\mathrm{device}} at frequencies of 6-10 GHz.

Extended results of the ST-FMR measurements of different oxide substrate/Py (5nm) samples and Si/SiO2_2/Pt (5nm)/Py (5nm) control sample.

We carried out the ST-FMR measurements with varying the frequency to quantify the spin-Hall ratio of Pt in control sample Si/SiO2_2/Pt (5nm)/Py (5nm). Figure S4 shows the ST-FMR spectra at frequencies of 5-10 GHz, and the symmetric components and antisymmetric components can be extracted via fitting by Eq. (1) in the main text (Figure S4). We obtained the effective demagnetization of Py by fitting the correlation between resonant frequency ff and resonant field H0H_0 via the Kittle formula

ST-FMR measurements of the Si/SiO2/Pt (5nm)/Py (5nm) control sample

Figure S4. ST-FMR measurements of the Si/SiO2_2/Pt (5nm)/Py (5nm) control sample. (a) ST-FMR spectra measured at frequencies of 5-10 GHz, φ=225∘\varphi= 225^\circ, P=15P = 15 dBm. (b) Symmetric and antisymmetric components extracted by the Lorentzian function fitting at frequencies of 5-10 GHz. (c) The resonant frequency ff as a function of the resonant

f=γ2πH0(H0+4πMeff)f = \frac{\gamma}{2\pi}\sqrt{H_0\left(H_0+4\pi M_{\mathrm{eff}}\right)}

, where γ=0.00294\gamma= 0.00294 GHz/Oe,

as shown in Figure S4. The spin-Hall ratio evaluated via

θSH=VSVA1+4πMeffH0eμ0Mstdℏ\theta_{\mathrm{SH}} = \frac{V_{\mathrm{S}}}{V_{\mathrm{A}}}\sqrt{1+\frac{4\pi M_{\mathrm{eff}}}{H_0}}\frac{e\mu_0 M_{\mathrm{s}}t_{\mathrm{d}}}{\hbar}

at frequencies of 5-10 GHz is shown in Figure S4 showing independent on frequency, and the spin-Hall ratio 0.071 ±\pm 0.002 is the mean value.

field H0H_0, which be well fit by the Kittle formula. (d) The spin-Hall ratio at frequencies of 5-10 GHz.

Symmetric and antisymmetric component of different oxide substrate/Py devices

Figure S5. Symmetric and antisymmetric component of (a)-(e) different oxide substrate/Py (5nm) devices extracted by the Lorentzian function fitting at frequencies of 5-10 GHz, P=15P = 15 dBm, and φ=225∘\varphi= 225^\circ.

Absence of damping-like effective field in different oxide-substrate/Py (5nm) samples from in-plane second harmonic Hall voltage measurements

We used the in-plane second harmonic Hall voltage measurements (SHHV), as shown in Figure S8, to confirm whether there exist damping-like torques in oxide substrate/Py samples. In the SHHV measurements, a 4mA-peak, 233Hz sine current is supplied by a current source, and the first and second harmonic Hall voltage responses are measured by two lock-in amplifiers, respectively. In the macro-spin approximation, the SHHV under an in-plane magnetic field is given by [1]

SHHV measurements of different oxide substrate/Py devices

Figure S8. SHHV measurements of different oxide substrate/Py devices. (a) Schematic of in-plane second harmonic Hall voltage measurement set-up. (b)-(e) The typical in-plane second harmonic Hall voltage responses under a 1000 Oe magnetic field of different oxide substrate/Py devices, which can be well fit by Eq. (8).

V2ω=Vacos⁡ϕ+Vpcos⁡2ϕcos⁡ϕ+VTsin⁡2ϕ(8)V_{2\omega}=V_a\cos\phi+V_p\cos2\phi\cos\phi+V_T\sin2\phi \tag*{(8)}

where ϕ\phi is the angle of HinH_{\mathrm{in}} with respect to the current direction, Va=VAHEHDL/2(Hin+Hk)+VANEV_a=V_{\mathrm{AHE}}H_{\mathrm{DL}}/2(H_{\mathrm{in}}+H_k)+V_{\mathrm{ANE}} is the contribution from damping-like field HDLH_{\mathrm{DL}}, and anomalous Nernst effect VANEV_{\mathrm{ANE}} due to the vertical thermal gradient,

VAHE=IRAHEV_{\mathrm{AHE}}=IR_{\mathrm{AHE}} is the anomalous Hall effect and RAHER_{\mathrm{AHE}} is the anomalous Hall effect resistance, Vp=VPHE(HFL+HOe)/HinV_p=V_{\mathrm{PHE}}(H_{\mathrm{FL}}+H_{\mathrm{Oe}})/H_{\mathrm{in}} is the contribution from field-like field HFLH_{\mathrm{FL}} and Oersted field HOeH_{\mathrm{Oe}}, VPHE=IRPHEV_{\mathrm{PHE}}=IR_{\mathrm{PHE}} is the planar Hall effect and RPHER_{\mathrm{PHE}} is the planar Hall effect resistance, VTV_T is the contribution from a thermal gradient along the direction of current [1]. Figure S8-(Figure S8) show the typical SHHV results of STO/Py, KTO/Py, Al2_2O3_3/Py, DSO/Py, LSAT/Py under a 1000 Oe magnetic field, which can be well fit by Eq. (8). To extract HDLH_{\mathrm{DL}}, we have determined RAHER_{AHE} of different oxide substrate/Py devices under the swept out-of-plane field HzH_z, as shown in Figure S9. Here, 4πMeff4\pi M_{eff} from the Kittle fitting in Figure S7 is employed to replace HkH_k. The linear correlation between VaV_a and −VAHE2(Hin+4πMeff)\frac{-V_{AHE}}{2(H_{in}+4\pi M_{eff})} is shown in Figure S10, and the main component VaV_a is mostly contributed by anomalous Nernst effect since VaV_a is hardly dependent of exerted magnetic field. The extracted damping-like fields HDLH_{DL} which are in the range of -0.18 ±\pm 0.21 Oe to 0.12 ±\pm 0.23 Oe for these five types of oxide-substrate/Py devices, as shown in Fig. 10(b), can be ignored within the error range.

Resistance of anomalous Hall effect under swept out-of-plane magnetic field

Figure S9. The resistance of anomalous Hall effect RAHER_{\mathrm{AHE}} under a swept of out-of-plane field HzH_z at range of -2 T to 2 T of different oxide substrate/Py Hall bars.

Damping-like effective field measurements for different oxide substrate/Py devices

Figure S10. (a) VaV_a as a function of −VAHE2(Hin+4πMeff)\frac{-V_{\mathrm{AHE}}}{2(H_{\mathrm{in}}+4\pi M_{\mathrm{eff}})}, with the slopes being damping-like effective field HDLH_{\mathrm{DL}}. (b) HDLH_{\mathrm{DL}} extracted from SHHV measurements of different oxide substrate/Py devices. Within the margin of error, HDLH_{\mathrm{DL}} of different oxide substrate/Py devices can be ignored.

The derivation of mixing voltage of the oxide substrate/Py samples from ST-FMR.

By solving the Landau-Lifshitz-Gilbert (LLG) equation

dm⃗dt=−γμ0m⃗×H⃗eff+αm⃗×dm⃗dt+τDLm⃗×(σ⃗×m⃗)+(τFL+γμ0H⃗RF)m⃗×σ⃗(9)\frac{d\vec{m}}{dt}=-\gamma\mu_0\vec{m}\times\vec{H}_{eff}+\alpha\vec{m}\times\frac{d\vec{m}}{dt}+\tau_{DL}\vec{m}\times\left(\vec{\sigma}\times\vec{m}\right)+\left(\tau_{FL}+\gamma\mu_0\vec{H}_{RF}\right)\vec{m}\times\vec{\sigma} \tag*{(9)}

where γ\gamma is the gyromagnetic ratio, α\alpha is the Gilbert damping coefficient, μ0\mu_0 is the permeability in vacuum, τDL\tau_{DL} is the damping-like torque, τFL\tau_{FL} is the field-like torque, H⃗eff\vec{H}_{eff} is the sum of external field H⃗ext\vec{H}_{ext} and the demagnetization field 4πMeff4\pi M_{eff}, H⃗RF\vec{H}_{RF} is the RF Oersted field, m⃗\vec{m} is the magnetization of FM layer, σ⃗\vec{\sigma} is the spin polarization, we can get the magnetization component of FM layer [2]:

my=12cos⁡φ2π(df/dHext)∣Hext=H0∣(τFL+γμ0HRF)1+(4πMeff/H0)+iτDLHext−H0+iΔH(10)m_y=\frac{1}{2}\frac{\cos\varphi}{2\pi\left(df/dH_{ext}\right)_{\left|H_{ext}=H_0\right|}}\frac{\left(\tau_{FL}+\gamma\mu_0H_{RF}\right)\sqrt{1+\left(4\pi M_{eff}/H_0\right)}+i\tau_{DL}}{H_{ext}-H_0+i\Delta H} \tag*{(10)}

where φ\varphi is the angle between RF current and external field, ff is the frequency of ST-FMR measurement, H0H_0 is the resonant field, MsM_s is the saturation magnetization, HkH_k is the magnetic anisotropy field, ΔH\Delta H is the linewidth.

The mixing voltage signal is

Vmix=12dRdφIRFRe⁡(my)(11)V_{mix}=\frac{1}{2}\frac{dR}{d\varphi}I_{RF}\operatorname{Re}\left(m_y\right) \tag*{(11)}

where IRFI_{RF} is the RF current through micro strip, RR is the resistance of microstrip.

For the oxide substrate/Py device, the leakage RF current IRF−substrateI_{RF-substrate} in the oxide substrate will generate an RF Oersted field HRF′H'_{RF}, as shown in Figure 3(a) in the main text. To evaluate the distribution of RF current in substrate, we did finite element simulations to investigate the distribution of RF current in the substrate. We focused solely on the RF current along the x-axis which generates an Oersted field along the y-axis, resulting in the production of an artifact symmetric signal. In our simulations, we introduced an RF power of 32 mW (~15 dBm) at 5 GHz. The material parameters utilized in the simulations are detailed in Table S1, and the simulation model is depicted in Figure S11.

ParameterValue
The resistivity of Pt electrode2.0×10−7 Ω⋅m2.0\times10^{-7}\ \Omega\cdot\mathrm{m}
The resistivity of Py4.5×10−7 Ω⋅m4.5\times10^{-7}\ \Omega\cdot\mathrm{m}
The relative permittivity of substrate10, 20, 100, 200, 300

Table S1. The material parameters used in simulations

Simulation model and typical simulation results of RF current distribution

Figure S11. (a) The simulation model of device used in our simulations. (b) The typical simulation results of the RF current distribution.

As shown in Figure S11, due to the spreading effect, the distribution of the xx-axis RF current in substrate is wider than the microstrip width, and the RF current is mostly distributed in the vicinity of the interface between the substrate and Py. According to the skin effect, employing 37% of the maximum RF current density as a threshold allows us to evaluate the spreading width of the xx-axis RF current distribution, as depicted in Figure S12. When using substrates with relative permittivity ranging from 10 to 300, the width of xx-axis RF current distribution is found to be 28.8μm to 29.4μm. With an order of magnitude increase in the relative permittivity of the substrate, there is a concomitant order of magnitude increase in the artifact signal, as shown in Figure 2. Consequently, we think that the primary factor affecting the magnitude of the artifact signal is the disparity in relative permittivity among different substrates.

Simulation results of RF current distributions

Figure S12. The simulation results of (a) the y-axis distribution of x-axis RF current in the interface between substrate and Py at x=0, z=0, and (b) the z-axis distribution of x-axis RF current in substrate at x=0, y=0.

Moreover, the impact of RF current flowing in the substrate outside the width of the microstrip on the Py layer should be disregarded, as the region where the y-axis component of the Oersted field is present extends beyond the width of the microstrip, and the z-axis component of the Oersted field generated by RF current flowing in substrate on either side of the microstrip cancels out each other. Therefore, the spreading effect of RF current in substrate could be ignored, and only the RF current flowing beneath the microstrip should be considered. As shown in Figure S12, the width of microstrip (∼18\sim18 μm) is one order larger than the skin depth (<1<1 μm) of RF current in substrate, so the Oersted field can be determined by the Ampere’s law approximately [3]. Considering the nonuniform distribution of leakage RF current in the skin depth, the Oersted field generated from RF current could be expressed in an integral form:

HRF′=∫−d0i∣JRF−substrate(z)∣ dz2=12i∫−d0∣IRF−substrate(z)∣ dzw dz=i2w∑∣IRF−substrate(z)∣(12)H^\prime_{\mathrm{RF}}=\int_{-d}^{0}\frac{i\lvert J_{\mathrm{RF-substrate}}(z)\rvert\,\mathrm{d}z}{2}=\frac{1}{2}i\int_{-d}^{0}\frac{\lvert I_{\mathrm{RF-substrate}}(z)\rvert\,\mathrm{d}z}{w\,\mathrm{d}z}=\frac{i}{2w}\sum\lvert I_{\mathrm{RF-substrate}}(z)\rvert \tag*{(12)}

where dd is the skin depth of RF current in substrate, ww is the width of microstrip, ∑IRF−substrate(z)\sum I_{\mathrm{RF-substrate}}(z) is the sum of RF current in the skin depth. In (12), ∑IRF−substrate(z)\sum I_{\mathrm{RF-substrate}}(z) can be replaced with ∣IRF−substrate∣\lvert I_{\mathrm{RF-substrate}}\rvert approximately, since the main distribution of RF current in substrate is in skin depth. Thus, HRF′H^\prime_{\mathrm{RF}} could be expressed approximately as (13):

HRF′=12wi∑∣IRF−substrate(z)∣≈i∣IRF−substrate∣2w.(13)H^\prime_{\mathrm{RF}}=\frac{1}{2w}i\sum\lvert I_{\mathrm{RF-substrate}}(z)\rvert\approx i\frac{\lvert I_{\mathrm{RF-substrate}}\rvert}{2w}. \tag*{(13)}

Since there is no spin source material in the oxide substrate/Py, the part associated with τDL\tau_{\mathrm{DL}} in Eq. (S3) could be deleted. Meanwhile, the field-like torque could also be ignored in oxide substrate/Py devices, which can be determined from the SHHV measurements as shown in Figure S8. Substituting Eq. (S6) into Eq. (S3), we can get:

my=12cos⁡φ2π(df/dHext)∣Hext=H0i1+(4πMeff/H0)⋅γμ0∣IRF−substrate∣2wHext−H0+iΔH(14)m_y=\frac{1}{2}\frac{\cos\varphi}{2\pi\left(\mathrm{d}f/\mathrm{d}H_{\mathrm{ext}}\right)_{\left\lvert H_{\mathrm{ext}}=H_0\right.}}\frac{i\sqrt{1+\left(4\pi M_{\mathrm{eff}}/H_0\right)}\cdot\gamma\mu_0\frac{\lvert I_{\mathrm{RF-substrate}}\rvert}{2w}}{H_{\mathrm{ext}}-H_0+i\Delta H} \tag*{(14)}

Furthermore, we can get the mixing voltage in oxide substrate/Py from (11) and (14):

Vmix′=−14dRdφ∣IRF∣cos⁡φΔH 2π(df/dHext)∣Hext=H01+(4πMeff/H0)γμ0∣IRF−substrate∣2wFS(Hext)=VS′FS(Hext)(15)V^\prime_{\mathrm{mix}}=-\frac{1}{4}\frac{\mathrm{d}R}{\mathrm{d}\varphi}\frac{\lvert I_{\mathrm{RF}}\rvert\cos\varphi}{\Delta H\,2\pi\left(\mathrm{d}f/\mathrm{d}H_{\mathrm{ext}}\right)_{\left\lvert H_{\mathrm{ext}}=H_0\right.}}\sqrt{1+\left(4\pi M_{\mathrm{eff}}/H_0\right)}\frac{\gamma\mu_0\lvert I_{\mathrm{RF-substrate}}\rvert}{2w}F_{\mathrm{S}}(H_{\mathrm{ext}})=V^\prime_{\mathrm{S}}F_{\mathrm{S}}(H_{\mathrm{ext}}) \tag*{(15)}

where FS(Hext)=ΔH2/[ΔH2+(Hext−H0)2]F_{\mathrm{S}}(H_{\mathrm{ext}})=\Delta H^2\big/\left[\Delta H^2+\left(H_{\mathrm{ext}}-H_0\right)^2\right] is the symmetric Lorentzian function.

ST-FMR and SHHV measurements of STO/Pt (5nm)/Py (5nm) and STO/Py (5nm)/Pt (5nm).

ST-FMR measurements of STO/Pt (5nm)/Py (5nm) and STO/Py (5nm)/Pt (5nm)

Figure S14. ST-FMR measurements of STO/Pt (5nm)/Py (5nm) and STO/Py (5nm)/Pt (5nm). Typical ST-FMR spectra measured at a frequency of 5 GHz, φ=225∘\varphi= 225^\circ, P=15P = 15 dBm of (a) STO/Pt/Py and (b) STO/Py/Pt, which can be well fit via Eq. (1) in the main text. Symmetric and antisymmetric components of (c) STO/Pt/Py and (d) STO/Py/Pt extracted by the Lorentzian function fitting at frequencies of 5–10 GHz. The Kittle formula fitting of resonant

frequency ff and the resonant field H0H_{0} of (e) STO/Pt/Py and (f) STO/Py/Pt.

Resistance of anomalous Hall effect under a swept out-of-plane field for STO/Pt/Py and STO/Py/Pt Hall bars

Figure S15. The resistance of anomalous Hall effect RAHER_{\mathrm{AHE}} under a swept of out-of-plane field HzH_{z} at range of -2 T to 2 T of (a) STO/Pt/Py and (b) STO/Py/Pt Hall bars.

Typical in-plane second harmonic Hall voltage responses for STO/Pt/Py and STO/Py/Pt

Figure S16. The typical in-plane second harmonic Hall voltage responses under a magnetic field of 1000 Oe of (a) STO/Pt/Py and (b) STO/Py/Pt, which are well fit by Eq. (1). During the SHHV measurements of STO/Pt/Py and STO/Py/Pt samples, an 8mA-peak, 233Hz sine current is supplied by a current source.

Spin-Hall ratio and damping-like effective field measurements

Figure S17. (a) The spin-Hall ratio θSH\theta_{\mathrm{SH}} of STO/Pt/Py and STO/Py/Pt evaluated θSH=VSVA1+4πMeffH0eμ0Mstdℏ\theta_{\mathrm{SH}}=\frac{V_{\mathrm{S}}}{V_{\mathrm{A}}}\sqrt{1+\frac{4\pi M_{\mathrm{eff}}}{H_{0}}}\frac{e\mu_{0}M_{\mathrm{s}}td}{\hbar} from ST-FMR measurements at frequencies of 5-10 GHz. (b) The linear dependence of VaV_{\mathrm{a}} on −VAHE/2(Hin+4πMeff)-V_{\mathrm{AHE}}/2(H_{\mathrm{in}}+4\pi M_{\mathrm{eff}}) of STO/Pt/Py and STO/Py/Pt, and the slopes being damping-like effective field HDLH_{\mathrm{DL}}.

ST-FMR results of the STO/SIO (17nm)/Py (5nm) sample.

ST-FMR spectra and symmetric and antisymmetric components

Figure S18. ST-FMR measurements of STO/SIO (17nm)/Py (5nm) sample. (a) Typical ST-FMR spectra of STO/SIO/Py at f=5f=5 GHz, φ=225∘\varphi=225^\circ, and P=15P=15 dBm. (b) Symmetric and antisymmetric components extracted by the Lorentzian functions fitting at frequencies of 5-10 GHz.

References

  1. [1]A. Manchon, J. Železný, I. M. Miron, T. Jungwirth, J. Sinova, A. Thiaville, K. Garello, and P. Gambardella, Current-induced spin-orbit torques in ferromagnetic and antiferromagnetic systems, Rev. Mod. Phys. 91, 035004 (2019).DOI
  2. [2]X. Han, X. Wang, C. Wan, G. Yu, and X. Lv, Spin-orbit torques: Materials, physics, and devices, Appl. Phys. Lett. 118, 120502 (2021).DOI
  3. [3]B. Dieny, I. L. Prejbeanu, K. Garello, P. Gambardella, P. Freitas, R. Lehndorff, W. Raberg, U. Ebels, S. O. Demokritov, J. Akerman, A. Deac, P. Pirro, C. Adelmann, A. Anane, A. V. Chumak, A. Hirohata, S. Mangin, S. O. Valenzuela, M. C. Onbaşlı, M. d’Aquino, G. Prenat, G. Finocchio, L. Lopez-Diaz, R. Chantrell, O. Chubykalo-Fesenko, and P. Bortolotti, Opportunities and challenges for spintronics in the microelectronics industry, Nat. Electron. 3, 446 (2020).arxiv.org/abs/1908.10584
  4. [4]I. M. Miron, K. Garello, G. Gaudin, P.-J. Zermatten, M. V. Costache, S. Auffret, S. Bandiera, B. Rodmacq, A. Schuhl, and P. Gambardella, Perpendicular switching of a single ferromagnetic layer induced by in-plane current injection, Nature 476, 189 (2011).DOI
  5. [5]K. Garello, F. Yasin, H. Hody, S. Couet, L. Souriau, S. H. Sharifi, J. Swerts, R. Carpenter, S. Rao, W. Kim, J. Wu, K. K. V. Sethu, M. Pak, N. Jossart, D. Crotti, A. Furnémont, and G. S. Kar, in 2019 Symposium on VLSI Circuits2019), pp. T194.DOI
  6. [6]Y. C. Wu, K. Garello, W. Kim, M. Gupta, M. Perumkunnil, V. Kateel, S. Couet, R. Carpenter, S. Rao, S. Van Beek, K. K. Vudya Sethu, F. Yasin, D. Crotti, and G. S. Kar, Voltage-Gate-Assisted Spin-Orbit-Torque Magnetic Random-Access Memory for High-Density and Low-Power Embedded Applications, Phys. Rev. Appl. 15, 064015 (2021).DOI
  7. [7]W. Cai, K. Shi, Y. Zhuo, D. Zhu, Y. Huang, J. Yin, K. Cao, Z. Wang, Z. Guo, Z. Wang, G. Wang, and W. Zhao, Sub-ns Field-Free Switching in Perpendicular Magnetic Tunnel Junctions by the Interplay of Spin Transfer and Orbit Torques, IEEE Electron. Device Lett. 42, 704 (2021).DOI
  8. [8]Q. Shao, P. Li, L. Liu, H. Yang, S. Fukami, A. Razavi, H. Wu, K. Wang, F. Freimuth, Y. Mokrousov, M. D. Stiles, S. Emori, A. Hoffmann, J. Åkerman, K. Roy, J. P. Wang, S. H. Yang, K. Garello, and W. Zhang, Roadmap of Spin-Orbit Torques, IEEE Trans. Magn. 57, 1 (2021).DOI
  9. [9]C. Safranski, J. Z. Sun, and A. D. Kent, A perspective on electrical generation of spin current for magnetic random access memories, Appl. Phys. Lett. 120, 160502 (2022).DOI
  10. [10]S. V. Aradhya, G. E. Rowlands, J. Oh, D. C. Ralph, and R. A. Buhrman, Nanosecond-Timescale Low Energy Switching of In-Plane Magnetic Tunnel Junctions through Dynamic Oersted-Field-Assisted Spin Hall Effect, Nano Lett. 16, 5987 (2016).pubmed.ncbi.nlm.nih.gov/27327619
  11. [11]M.-H. Nguyen and C.-F. Pai, Spin-orbit torque characterization in a nutshell, APL Mater. 9, 030902 (2021).DOI
  12. [12]L. Liu, T. Moriyama, D. C. Ralph, and R. A. Buhrman, Spin-Torque Ferromagnetic Resonance Induced by the Spin Hall Effect, Phys. Rev. Lett. 106, 036601 (2011).DOI
  13. [13]T. Nan, S. Emori, C. T. Boone, X. Wang, T. M. Oxholm, J. G. Jones, B. M. Howe, G. J. Brown, and N. X. Sun, Comparison of spin-orbit torques and spin pumping across NiFe/Pt and NiFe/Cu/Pt interfaces, Phys. Rev. B 91, 214416 (2015).DOI
  14. [14]T. Nan, C. X. Quintela, J. Irwin, G. Gurung, D. F. Shao, J. Gibbons, N. Campbell, K. Song, S. Y. Choi, L. Guo, R. D. Johnson, P. Manuel, R. V. Chopdekar, I. Hallsteinsen, T. Tybell, P. J. Ryan, J. W. Kim, Y. Choi, P. G. Radaelli, D. C. Ralph, E. Y. Tsymbal, M. S. Rzchowski, and C. B. Eom, Controlling spin current polarization through non-collinear antiferromagnetism, Nat. Commun. 11, 4671 (2020).DOI
  15. [15]T. Nan, T. J. Anderson, J. Gibbons, K. Hwang, N. Campbell, H. Zhou, Y. Q. Dong, G. Y. Kim, D. F. Shao, T. R. Paudel, N. Reynolds, X. J. Wang, N. X. Sun, E. Y. Tsymbal, S. Y. Choi, M. S. Rzchowski, Y. B. Kim, D. C. Ralph, and C. B. Eom, Anisotropic spin-orbit torque generation in epitaxial SrIrO₃ by symmetry design, Proc. Natl. Acad. Sci. U.S.A. 116, 16186 (2019).
  16. [16]A. R. Mellnik, J. S. Lee, A. Richardella, J. L. Grab, P. J. Mintun, M. H. Fischer, A. Vaezi, A. Manchon, E. A. Kim, N. Samarth, and D. C. Ralph, Spin-transfer torque generated by a topological insulator, Nature 511, 449 (2014).DOI
  17. [17]Y. Wang, P. Deorani, K. Banerjee, N. Koirala, M. Brahlek, S. Oh, and H. Yang, Topological Surface States Originated Spin-Orbit Torques in Bi₂Se₃, Phys. Rev. Lett. 114, 257202 (2015).arxiv.org/abs/1505.07937
  18. [18]M. Hayashi, J. Kim, M. Yamanouchi, and H. Ohno, Quantitative characterization of the spin-orbit torque using harmonic Hall voltage measurements, Phys. Rev. B 89, 144425 (2014).DOI
  19. [19]U. H. Pi, K. Won Kim, J. Y. Bae, S. C. Lee, Y. J. Cho, K. S. Kim, and S. Seo, Tilting of the spin orientation induced by Rashba effect in ferromagnetic metal layer, Appl. Phys. Lett. 97, 162507 (2010).DOI
  20. [20]L. Zhu, L. Zhu, and R. A. Buhrman, Fully Spin-Transparent Magnetic Interfaces Enabled by the Insertion of a Thin Paramagnetic NiO Layer, Phys. Rev. Lett. 126, 107204 (2021).DOI
  21. [21]I. H. Kao, R. Muzzio, H. Zhang, M. Zhu, J. Gobbo, S. Yuan, D. Weber, R. Rao, J. Li, J. H. Edgar, J. E. Goldberger, J. Yan, D. G. Mandrus, J. Hwang, R. Cheng, J. Katoch, and S. Singh, Deterministic switching of a perpendicularly polarized magnet using unconventional spin-orbit torques in WTe₂, Nat. Mater. 21, 1029 (2022).
  22. [22]H. Wu, S. A. Razavi, Q. Shao, X. Li, K. L. Wong, Y. Liu, G. Yin, and K. L. Wang, Spin-orbit torque from a ferromagnetic metal, Phys. Rev. B 99, 184403 (2019).DOI
  23. [23]C.-F. Pai, M. Mann, A. J. Tan, and G. S. D. Beach, Determination of spin torque efficiencies in heterostructures with perpendicular magnetic anisotropy, Phys. Rev. B 93, 144409 (2016).
  24. [24]L. Zhu, D. C. Ralph, and R. A. Buhrman, Lack of Simple Correlation between Switching Current Density and Spin-Orbit-Torque Efficiency of Perpendicularly Magnetized Spin-Current-Generator-Ferromagnet Heterostructures, Phys. Rev. Appl. 15, 024059 (2021).
  25. [25]T. Xing, C. Zhou, C. X. Wang, Z. Li, A. N. Cao, W. L. Cai, X. Y. Zhang, B. Ji, T. Lin, Y. Z. Wu, N. Lei, Y. G. Zhang, and W. S. Zhao, Direct detection of spin-orbit effective fields through magneto-optical Kerr effect, Phys. Rev. B 101, 224407 (2020).DOI
  26. [26]M. Montazeri, P. Upadhyaya, M. C. Onbasli, G. Yu, K. L. Wong, M. Lang, Y. Fan, X. Li, P. Khalili Amiri, R. N. Schwartz, C. A. Ross, and K. L. Wang, Magneto-optical investigation of spin-orbit torques in metallic and insulating magnetic heterostructures, Nat. Commun. 6, 8958 (2015).DOI
  27. [27]B. Grover, B. K. Hazra, T. Ma, B. Pal, N. Bernstein, A. Rothschild, A. K. Srivastava, S. Choudhury, G. Woltersdorf, A. Capua, and S. S. P. Parkin, Crystallographic dependence of the spin Hall angle in epitaxial Pt films: Comparison of optical and electrical detection of spin-torque ferromagnetic resonance techniques, Appl. Phys. Lett. 120, 172406 (2022).DOI
  28. [28]X. Huang, S. Sayed, J. Mittelstaedt, S. Susarla, S. Karimed­diny, L. Caretta, H. Zhang, V. A. Stoica, T. Gosavi, F. Mahfouzi, Q. Sun, P. Ercius, N. Kioussis, S. Salahuddin, D. C. Ralph, and R. Ramesh, Novel Spin-Orbit Torque Generation at Room Temperature in an All-Oxide Epitaxial La₀.₇Sr₀.₃MnO₃/SrIrO₃ System, Adv. Mater. 33, 2008269 (2021).DOI
  29. [29]J. Zhou, X. Shu, W. Lin, D. F. Shao, S. Chen, L. Liu, P. Yang, E. Y. Tsymbal, and J. Chen, Modulation of Spin-Orbit Torque from SrRuO₃ by Epitaxial-Strain-Induced Octahedral Rotation, Adv. Mater. 33, 2007114 (2021).DOI
  30. [30]J. Wei, H. Zhong, J. Liu, X. Wang, F. Meng, H. Xu, Y. Liu, X. Luo, Q. Zhang, Y. Guang, J. Feng, J. Zhang, L. Yang, C. Ge, L. Gu, K. Jin, G. Yu, and X. Han, Enhancement of Spin-Orbit Torque by Strain Engineering in SrRuO₃ Films, Adv. Funct. Mater. 31, 2100380 (2021).
  31. [31]S. Karimeddiny, J. A. Mittlestaedt, R. A. Buhrman, and D. C. Ralph, Transverse and Longitudinal Spin-Torque Ferromagnetic Resonance for Improved Measurement of Spin-Orbit Torque, Phys. Rev. Appl. 14, 024024 (2020).DOI
  32. [32]T. M. Cham, S. Karimeddiny, V. Gupta, J. A. Mittlestaedt, and D. C. Ralph, Separation of Artifacts from Spin-Torque Ferromagnetic Resonance Measurements of Spin-Orbit Torque for the Low-Symmetry Van der Waals Semi-Metal ZrTe₃, Adv. Quantum Technol. 5, 2100111 (2022).
  33. [33]J. Zhang, J. Zhang, X. Chi, R. Hao, W. Chen, H. Yang, D. Zhu, Q. Zhang, W. Zhao, H. Zhang, and J. Sun, Giant efficiency for charge-to-spin conversion via the electron gas at the LaTiO₃/SrTiO₃ interface, Phys. Rev. B 105, 195110 (2022).DOI
  34. [34]H. Chen and D. Yi, Spin–charge conversion in transition metal oxides, APL Mater. 9, 060908 (2021).DOI
  35. [35]A. Bose, J. N. Nelson, X. S. Zhang, P. Jadaun, R. Jain, D. G. Schlom, D. C. Ralph, D. A. Muller, K. M. Shen, and R. A. Buhrman, Effects of Anisotropic Strain on Spin–Orbit Torque Produced by the Dirac Nodal Line Semimetal IrO₂, ACS Appl. Mater. Interfaces 12, 55411 (2020).DOI
  36. [36]Y. Ou, Z. Wang, C. S. Chang, H. P. Nair, H. Paik, N. Reynolds, D. C. Ralph, D. A. Muller, D. G. Schlom, and R. A. Buhrman, Exceptionally High, Strongly Temperature Dependent, Spin Hall Conductivity of SrRuO₃, Nano Lett. 19, 3663 (2019).DOI
  37. [37]S. Hori, K. Ueda, T. Kida, M. Hagiwara, and J. Matsuno, Spin–orbit torque generation in bilayers composed of CoFeB and epitaxial SrIrO₃ grown on an orthorhombic DyScO₃ substrate, Appl. Phys. Lett. 121, 022402 (2022).DOI
  38. [38]H. Zhang, Z. Zhu, Y. Zhu, X. Chen, Q. Jiang, J. Wei, C. Zhao, J. Zhang, F. Han, H. Yang, D. Zhu, H. Wu, Y. Chen, F. Hu, B. Shen, W. Zhao, J. Zhang, G. Yu, and J. Sun, Fermi-Level-Dependent Charge-to-Spin Conversion of the Two-Dimensional Electron Gas at the γ-Al₂O₃/KTaO₃ Interface, Phys. Rev. Appl. 19, 034045 (2023).DOI
  39. [39]E. Lesne, Y. Fu, S. Oyarzun, J. C. Rojas-Sánchez, D. C. Vaz, H. Naganuma, G. Sicoli, J. P. Attané, M. Jamet, E. Jacquet, J. M. George, A. Barthélémy, H. Jaffrès, A. Fert, M. Bibes, and L. Vila, Highly efficient and tunable spin-to-charge conversion through Rashba coupling at oxide interfaces, Nat. Mater. 15, 1261 (2016).DOI
  40. [40]H. Zhang, Y. Ma, H. Zhang, X. Chen, S. Wang, G. Li, Y. Yun, X. Yan, Y. Chen, F. Hu, J. Cai, B. Shen, W. Han, and J. Sun, Thermal Spin Injection and Inverse Edelstein Effect of the Two-Dimensional Electron Gas at EuO-KTaO₃ Interfaces, Nano Lett. 19, 1605 (2019).DOI
  41. [41]R. C. Neville, B. Hoeneisen, and C. A. Mead, Permittivity of Strontium Titanate, J. Appl. Phys. 43, 2124 (1972).DOI
  42. [42]V. Trepakov, V. Vikhinin, M. Savinov, P. Syrnikov, S. Kapphan, V. Lemanov, H. Hesse, and L. Jastrabik, Dielectric permittivity and Fe-and Cu-doping effect in KTaO₃ and K₁₋ₓLiₓTaO₃, Ferroelectrics 235, 59 (1999).DOI
  43. [43]J. Sinova, S. O. Valenzuela, J. Wunderlich, C. H. Back, and T. Jungwirth, Spin Hall effects, Rev. Mod. Phys. 87, 1213 (2015).
  44. [44]V. M. Edelstein, Spin polarization of conduction electrons induced by electric current in two-dimensional asymmetric electron systems, Solid State Commun. 73, 233 (1990).DOI
  45. [45]D. MacNeill, G. M. Stiehl, M. H. D. Guimaraes, R. A. Buhrman, J. Park, and D. C. Ralph, Control of spin-orbit torques through crystal symmetry in WTe₂/ferromagnet bilayers, Nat. Phys. 13, 300 (2017).
  46. [46]Y. Wang, P. Deorani, X. Qiu, J. H. Kwon, and H. Yang, Determination of intrinsic spin Hall angle in Pt, Appl. Phys. Lett. 105, 152412 (2014).DOI
  47. [47]Y. Fan, H. Li, M. Dc, T. Peterson, J. Held, P. Sahu, J. Chen, D. Zhang, A. Mkhoyan, and J.-P. Wang, Spin pumping and large field-like torque at room temperature in sputtered amorphous WTe₂−ₓ films, APL Mater. 8, 041102 (2020).arxiv.org/abs/2002.09510
  48. [48]K. Kondou, H. Chen, T. Tomita, M. Ikhlas, T. Higo, A. H. MacDonald, S. Nakatsuji, and Y. Otani, Giant field-like torque by the out-of-plane magnetic spin Hall effect in a topological antiferromagnet, Nat. Commun. 12, 6491 (2021).DOI
  49. [49]See Supplementary Material at [URL] for more details about the measurements of impedance , evaluation of spin-Hall ratio of Si/SiO₂/Pt (5nm)/Py (5nm) sample, ST-FMR and SHHV results of different oxide substrate/Py devices, the derivation of ST-FMR mixing voltage of oxide substrate/Py device, ST-FMR and SHHV results of STO/Pt (5nm)/Py (5nm) and STO/Py (5nm)/Pt (5nm) samples, and ST-FMR results of STO/SIO (17nm)/Py (5nm) sample.
  50. [50]C. Zuccaro, I. Ghosh, K. Urban, N. Klein, S. Penn, and N. M. Alford, Materials for HTS-shielded dielectric resonators, IEEE Trans. Appl. Supercond. 7, 3715 (1997).DOI
  51. [51]V. Bovtun, M. Kempa, S. Kamba, V. Pashkov, V. Molchanov, Y. Poplavko, and Y. Yakymenko, in IEEE XXXIII International Scientific Conference Electronics and Nanotechnology (ELNANO) (2013), pp. 17-20.
  52. [52]H. Birey, Thickness dependence of the dielectric constant and resistance of Al₂O₃ films, J. Appl. Phys. 48, 5209 (1977).DOI
  53. [53]S. Emori, T. Nan, A. M. Belkessam, X. Wang, A. D. Matyushov, C. J. Babroski, Y. Gao, H. Lin, and N. X. Sun, Interfacial spin-orbit torque without bulk spin-orbit coupling, Phys. Rev. B 93, 180402(R) (2016).DOI
  54. [54]Seki, Takeshi and Lau, Yong-Chang and Iihama, Satoshi and Takanashi, Koki, Spin-orbit torque in a Ni-Fe single layer, Phys. Rev. B 104, 094430 (2021).DOI
  55. [55]Tsukahara, Ayaka and Ando, Yuichiro and Kitamura, Yuta and Emoto, Hiroyuki and Shikoh, Eiji and Delmo, Michael P. and Shino, Teruya and Shiraishi, Masashi, Self-induced inverse spin Hall effect in permalloy at room temperature, Phys. Rev. B 89, 235317 (2014).DOI
  56. [56]L. M. Vicente-Arche, J. Bréhin, S. Varotto, M. Cosset-Cheneau, S. Mallik, R. Salazar, P. Noël, D. C. Vaz, F. Trier, S. Bhattacharya, A. Sander, P. Le Fèvre, F. Bertran, G. Saiz, G. Ménard, N. Bergeal, A. Barthélémy, H. Li, C.-C. Lin, D. E. Nikonov, I. A. Young, J. E. Rault, L. Vila, J.-P. Attané, and M. Bibes, Spin-Charge Interconversion in KTaO₃ 2D Electron Gases, Adv. Mater. 33, 2102102 (2021).DOI
  57. [57]A. Tang, T. Xu, S. Liu, Y. Liang, H. Chen, D. Yan, Y. Shi, P. Yu, R. Yu, and Y. Lin, Implementing Complex Oxides for Efficient Room-Temperature Spin–Orbit Torque Switching, Adv. Electron. Mater. 8, 2200514 (2022).DOI
  58. [58]D. Zheng, J. Lan, B. Fang, Y. Li, C. Liu, J. O. Ledesma-Martin, Y. Wen, P. Li, C. Zhang, Y. Ma, Z. Qiu, K. Liu, A. Manchon, and X. Zhang, High-Efficiency Magnon-Mediated Magnetization Switching in All-Oxide Heterostructures with Perpendicular Magnetic Anisotropy, Adv. Mater. 34, 2203038 (2022).DOI

Paper details

vlnc:W_mjvj384u
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Physics and Astronomy

Not recorded for this paper

View statement
Looking for related papers…
Contents